WAYNE C. 1SPHORDING Department of Geology, University of South , Mobile, GEORGE M. LAMB Alabama 36608 Age and Origin of the Citronelle Formation in Alabama

ABSTRACT early and are, in turn, overlain by Pleistocene-age terrace and high terrace The sediments forming the Citronelle For- deposits. mation occur as a nearly continuous blanket The thickness of the formation cannot be along the southern margin of the Coastal determined with certainty because reworking Plain and can be traced from Texas across the has incorporated the materials into overlying Gulf Coast into peninsular . The lack Pleistocene deposits at many locations and of evidence, however, has made this the two units cannot be differentiated. In formation the subject of a heated age con- addition, where the Citronelle is the upper- troversy since it was first described in 1916. most unit, recent erosion has developed con- At present, its age is given by various investi- siderable relief on the formation, complicating gators as late Miocene, Pliocene, Plio- thickness measurments. Best estimates place Pleistocene or Pleistocene. the maximum thickness at approximately 300 The recent discovery of abundant verte- ft near the Alabama- border. Most brate , at a site in northern Mobile other locations have sections ranging from County, Alabama, near the base of the forma- a few tens to 150 ft. tion, indicates that the maximum age of the Citronelle is mid-Pliocene (Hemphillian). Origin Other evidence is present to indicate that the The origin of the Citronelle Formation is rocks of fossil zone represent a brackish es- as controversial as the age. It has been con- tuary that was later filled by river encroach- sidered, variously, as (l) fluvatile terrace de- ment, forming the overlying fluviatile sedi- posits (Fisk, 1938; Alt and Brooks, 1965); ments typical of the Citronelle Formation. (2) Pleistocene glacial "southern drift" Hence, the faunal evidence, coupled with (Hilgard, 1866); (3) marine deposits (McGee, existing pollen data from the upper sediments 1891; Harris and Veatch, 1899); (4) transi- of the Citronelle in nearby Florida, now in- tional marine deposits (Matson, 1916); and dicates that deposition of this formation (5) as deposits of pre-glacial coalescing, began in the middle Pliocene and continued braiding streams. The latter view was held into the pre-Nebraskan Pleistocene. by Clendenin (1896) and Deering (1958) and has since been supported by Rosen (1969) DISCUSSION who conducted an excellent detailed study on the mineralogy and texture of Citronelle Introduction sediments in . One of the major controversies existing in the literature of the Coastal Plain of the Gulf Age Problem Coast involves the age of the widespread The sediments of the Citronelle Formation Citronelle Formation. The sands, gravels, and were originally included as part of the no clays that make up this unit can be traced, longer recognized "Lafayette Formation" and almost continuously, from Florida westward were not given individual formation status through Alabama, Mississippi, Louisiana, until 1916 when Matson proposed the name and into Texas, and in many places they cap for the thick sands and gravels near the town the higher hills near the southern margin of of Citronelle in northern Mobile County, the Gulf Coastal Plain. Stratigraphically, these Alabama. Later that same year, Berry (1916) sediments rest unconformably on formations discovered plant fossils in a clay bed at an that range in age from late Oligocene to exposure on the Gulf, Mobile and Ohio

Geological Society of America Bulletin, v. 82, p. 775-780, 1 fig., March 1971 775

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/82/3/775/3432985/i0016-7606-82-3-775.pdf by guest on 25 September 2021 776 ISPHORDING AND LAMB-CITRONELLE FORMATION IN ALABAMA

Railroad right-of-way 6 mi south of Citronelle, former had ever been found in the Citronelle and on the basis of the contained flora, con- and the few invertebrates reported were not firmed the Pliocene age assigned by Matson diagnostic species. (1916). Later investigators have argued that the clay bed in which the flora was found MOBILE COUNTY FOSSIL actually lay beneath the Citronelle Formation LOCALITY and was separated from it by either a fault (Roy, 1939) or disconformity (Carlston, Description and Age 1950). Stringfield and LaMoreaux (1957) A dark, gray, carbonaceous silty clay and believed that any unconformity present was clayey sand, approximately 20 ft thick was minor, and that this did not preclude the discovered at the base of the Citronelle For- plants belonging to the Citronelle Forma- mation on Chickasabogue Creek, in the tion. They also noted that a similar leaf- NW'4 sec. 27, T. 2 S., R. 2 W., in northern bearing clay bed was present at Red Bluff, Mobile County. This clay bed contains an on Perdido Bay west of Pensacola, Florida, abundance of carbonized wood fragments and was underlain by sediments of typical and also contains a several foot thick interval Citronelle lithology.(The senior author visited near the base with abundant vertebrate re- Berry's locality where the plant fossils were mains. The fossils so far obtained have been found and was unable to find definite evi- identified by Dr. Frank C. Whitmore, Jr., of dence of either a fault or unconformity. The the Geological Survey and the sands overlying the clay horizon did appear United States National Museum, with help- to have been subjected to some penecon- ful discussions from S. David Webb, Thomas temporaneous deformation, but this died out A. Patton, Beryl Taylor, and Morris F. Skinner laterally after a short distance). Doering aiding Dr. Whitmore in his identifications. (1958), however, questioned the age of the Whitmore (1970, oral commun.) lists the flora described by Berry and concluded that following vertebrate fossils found in the there was no valid reason to assign a Pliocene Chickasabogue Creek locality: age to the flora and that an early Pleistocene age was more logical. Marsh (1964) reports Osteichthyes that this conclusion was supported by pollen . Mackerels and related fish. Reptillia studies carried out by Estella Leopold of the Chelonia U.S. Geological Survey, who reported that sp. Soft-shelled turtle, either spinifer the flora present (p. 83) "... provides clear otferox species (Dart, 1970 written commun). fossil evidence of a Quaternary age for the Crocodilia. Inarticulated plates. middle and upper parts of the Citronelle For- Mammalia mation in westernmost Florida." More Perissodactyla recently, however, Alt and Brooks (1965) Hipparion phosphorum Simpson. Medium- have concluded, on the basis of soil associa- sized grazing horse. tions and distribution of major terrace Nannippus cf. N. lenticularis (Cope). Small, swift, grazing horse. deposits, that the Citronelle Formation in sp. Short-legged rhinoceros; peninsular Florida probably deserves a late probably amphibious. Miocene age. They (p. 408) "... find no Artiodactyla evidence to favor the Pleistocene date pro- cf. S. tricornatusStitton. Prong- posed by numerous authors . . .". Similarly, horn-like browser. a late Miocene age had earlier been proposed Camelid for the Citronelle sediments of the Lake Wales Cervid Ridge in Florida by Ketner and McGreevy Cetacea Pomatodelphis inaequalis Allen. Long-beaked (1959); however, they questioned whether river dolphin. there were any "true" Citronelle sediments in the Florida peninsula. Hence, at the time He further says that on the basis of this of this writing, the age of the Citronelle fauna, the age of the deposit is Hemphillian Formation is given, variously, as late Mio- (middle Pliocene) and the closest affinities cene, Pliocene, or as early Pleistocene. The of the fauna are with the Bone Valley Forma- age problem can be reduced simply to a lack tion of central Florida. This fossil site is of vertebrate or invertebrate fossils on which especially important in that it is the only a reliable date may be placed. None of the known location in the central Gulf Coast

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/82/3/775/3432985/i0016-7606-82-3-775.pdf by guest on 25 September 2021 MOBILE COUNTY FOSSIL LOCALITY 777

that contains a vertebrate fauna of this age. ture (that is, clayey sands) and wood frag- Hence, it may provide long sought-after in- ments become especially abundant. The lower formation about species distribution and part of Unit 1 is exposed approximately 100 migration of vertebrates between other known yds downstream from the fossil site and the locations in Florida and Texas. Further clay was found to contain some poorly pre- attempts by a University of South Alabama served molds of pelecypods, thecamoebids and Smithsonian Institution group to exca- and possible members of the foraminiferal vate this locality fully have been temporarily genus Trochammina. These all disappear up- thwarted, however, by a somewhat recalcitrant ward in the section with no evidence of any landowner. microfossils or other invertebrates in the ver- tebrate bed (Unit 2) or overlying units. This Paleoenvironment change would lend credence to the idea of A detailed micro-faunal, mineralogical and environmental change from brackish estu- textural analysis of the clay bed determined arine or marsh conditions to purely fluvial that it represents sediments deposited in a conditions with time. The nature of the verte- lagoon, estuary, or marsh that was receiving brate fossils suggests an environment not enough run-off so the water had low salinity. unlike the present-day delta of the Mobile Humid tropical or sub-tropical climatic con- River system. ditions are believed to have existed in the This conclusion is also supported by x-ray central Gulf Coast at this time (see Isphording, analysis of the clay-sized minerals present. 1970) and would explain the increased run- The minerals in Unit 1 consist largely of off and brackishness of marginal marine montmorillonite (77 percent) with lesser swamps and marshes. The more typical sands, amounts of kaolinite (12 percent), illite (6 gravels and clays of the Citronelle Formation percent), and quartz (5 percent). Upward in that overlie the clay bed are thought to be the section, montmorillonite gradually de- the product of increased fluvial activity along creases until in Unit 5 it makes up only 60 the southern margin of the Coastal Plain percent of the sample; kaolinite increases that resulted from epeirogenic uplift of the upward in the section and totals 25 percent continental interior at this time. in Unit 5. This change is not thought to be a The lowermost portion of the clay bed result of alteration of the 14 angstrom min- exposed (Unit 1, Fig. 1) is silty clay to clayey erals to kaolinite but rather only to reflect silt and contains relatively few fragments of the gradual decrease in the accumulation of carbonized wood. Upward in the section the marine or transitional marine sediments re- sediments become increasingly coarse in tex- sulting from the pro-grading of the marsh and river encroachment. The detrital minerals present also merit special attention as this is the only location so far encountered in the central and eastern Gulf Coast where epidote, hornblende and garnet have been found in greater than trace amounts in the Citronelle Formation (see Table l). In Units 1 through 3 these minerals combined to form 5 to 7 percent of the heavy mineral fraction (125 to 250 micron interval) and their presence is significant for several reasons: (l) they are not encountered in typical Citronelle sediments in either Florida or Alabama but are present, in small amounts, in those from Louisiana (see Rosen, 1969); (2) they are not found in the Miocene sedi- ments of Alabama but are very abundant in Florida Miocene strata; and (3) they are not found in the sediments of the Pliocene Graham Ferry Formation nearby in Mississippi, the Figure 1. Stratigraphic section at the Mobile only other formation to which the clay bed County fossil locality. might have been assigned. Although their

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/82/3/775/3432985/i0016-7606-82-3-775.pdf by guest on 25 September 2021 778 ISPHORDING AND LAMB-CITRONELLE FORMATION IN ALABAMA

TABLE 1. HEAVY MINERAL ANALYSES FROM THE QTRONELLE FORMATION IN ALABAMA AND FLORIDA

Type Section Clermont Sand Grandin Sand Yniestra, Mobile Co. Citronelle, Mine, Lake Co., Mine, Putnam Co., Escambia Co., Fossil Locality Ala. Fla.1 Fla.1 Fla.2 Unit 1 Unit 3 Unit 6 Epidote Absent to VR R R Garnet R R Hornblende R R Kyanite C 3 3 C C C VC Sillimanite VR VC VC R VR Rutile R R C C R R C Stautolite C R R C C C C Tourmaline C R R C C R C Zircon R C C C R C C Ilmenite A VA VA A VC A VC Leucoxene C C C C A VC VC 1 modified after Pirkle and others, (1965) F = flood (>75%) 2 modified after Martens (1970, written commun.) VA = very abundant (50-75%) 3 combined sillimanite and kyanite A = abundant (25-50%) VC = very common (15-25%) C = common (5-15%) R = rare (1-5%) VR = very rare (<1%)

hornblende, epidote and garnet content and represent materials deposited in transitional their clay mineralogy are remarkably similar marine environments. They are overlain by to Miocene sediments in western Florida, the more typical aggradation deposits of these sediments differ from those in Florida streams that form the bulk of the Citronelle in that the mineral sillimanite is scarce to Formation in the Gulf Coast. Vertebrate fossil absent, whereas in Florida Miocene sedi- evidence indicates that deposition of the ments, it commonly occurs in amounts rang- Citronelle sediments commenced in the ing up to 25 percent of the 62-125 micron middle Pliocene and, based on pollen data heavy mineral fraction (see Pirkle and others, from nearby sediments in western Florida, 1965; Isphording, 1970; Pirkle and Yoho, continued into the pre-Nebraskan Pleisto- 1970). It should also be noted that even cene. The age of the Citronelle Formation, in though sillimanite is virtually lacking in the the central Gulf Coast, can thus be established sediments from the clay horizon and typical as mid-Pliocene through early Pleistocene. Citronelle sediments in Alabama, it forms a The age problem itself, however, will con- major part of the Citronelle heavy mineral tinue to persist in many areas of the eastern fraction in samples in peninsular Florida. and western Gulf Coast because of the mis- This strongly suggests that sediments de- use of the term "Citronelle Formation." scribed as belonging to the Citronelle For- Deposits described as belonging to the mation in the Gulf Coast are the products of Citronelle Formation in Florida have, in some several different source areas and makes cor- cases, been established as being younger (or relation on the basis of mineralogy, except in older) than the sediments described at the local areas, hazardous at best. This can be fossil locality and nearby type section of the especially well seen at the Mobile County formation at Citronelle, Alabama. This is fossil locality where hornblende, epidote and particularly true in the case of the sediments garnet are found in Units 1 through 3 but are that form the Lake Wales Ridge in the central not present in Units 4 and 5, or in the typical Florida peninsula. If Ketner and McGreevy Citronelle sands that make up Unit 6. This (1959) and Alt and Brooks (1965) are correct indicates that the source for these minerals in their conclusion that these sediments are had been depleted by the time the upper late Miocene in age, then they cannot belong units were deposited. to the Citronelle Formation. These sediments should be re-defined and assigned to a new SUMMARY AND CONCLUSIONS or appropriate stratigraphic unit and the name The basal Citronelle sediments at the ver- "Citronelle" should no longer be used when tebrate fossil site in northern Mobile County referring to them. Similarly, many terrace

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/82/3/775/3432985/i0016-7606-82-3-775.pdf by guest on 25 September 2021 REFERENCES CITED 779

deposits of Pleistocene age in Mississippi, Fisk, H. N. Geology of Grant and LaSalle Alabama, and the western panhandle of Parishes: La. Geol. Surv., Bull., Vol. 10, Florida, are also included in the Citronelle 246 p., 1938. Harris, G. D.; and Veatch, A. C. A prelimi- Formation but may not actually belong to nary report on the geology of Louisiana: this unit. Hopefully, now that the age of the La. State Univ., State Exp. Sta. Rep., 138 formation has been established, all such p., 1899. questionable deposits can be re-examined Hilgard, E. W. On the Quaternary formations and properly assigned, and the term ".Citron- of the state of Mississippi: Amer. J. Sci., elle Formation" will no longer conjure up Vol. 91, p. 311-325, 1866. the nebulous picture among Gulf Coast geol- Isphording, W. C. Late Tertiary paleoclimate ogists that it has in the past. of eastern United States: Amer. Ass. Petrol. Geol., Bull., Vol. 54, p. 334-343, 1970. ACKNOWLEDGMENTS Ketner, K. B.; and McGreevy, L. J. Stratigra- phy of the area between Hernando and We would like to take this opportunity to Hardee Counties, Florida: U.S. Geol. Surv., thank Jerry W. Wooten, William C. Moody, Bull. 1074-C, p. 49-124, 1959. Bruce L. Isphording, and Ronald F. Turner Marsh, O. T. Geology of Escambia and Santa for their assistance in collecting the samples Rosa Counties, western Florida panhandle: used in this study. Dr. Frank C. Whitmore, Fla. Geol. Surv., Bull. 46, 140 p., 1964. of the United States Geological Survey and Maison, G. C. The Pliocene Citronelle Forma- United States National Museum, merits spe- tion of the Gulf Coastal Plain: U.S. Geol. cial thanks for visiting the site, identifying Surv., Prof. Pap. 98, p. 167-192, 1916. McGee, W. J. The Lafayette Formation: U.S. the fauna, and for his encouragement through- Geol. Surv., 12th Ann. Rep., p. 327-521, out the study. The authors are also grateful 1891. for the financial assistance provided by the Pirkle, E. C.; Yoho, W. H.; and Allen, A. T. Research Committee, University of South Hawthorne, Bone Valley and Citronelle Alabama. sediments of Florida: Fla. Acad. Sci., Quart J., Vol. 28, p. 7-58, 1965. Pirkle, E. C.; and Yoho, W. H. The heavy REFERENCES CITED mineral ore body of Trail Ridge, Florida: Alt, D.; and Brooks, H. K. Age of the Florida Econ. Geol., Vol. 65, p. 17-30, 1970. marine terraces: J. Geol., Vol. 73, p. 406- Rosen, N. C. Heavy minerals and size analysis 411, 1965. of the Citronelle Formation of the Gulf Berry, E. W. The flora of the Citronelle Forma- Coastal Plain: J. Sediment. Petrology, tion: U.S. Geol. Surv., Prof. Pap. 98, p. Vol. 39, p. 1552-1565, 1969. 193-208, 1916. Roy, C. J. Type locality of the Citronelle For- Carlston, C. W. Pleistocene history of coastal mation, Citronelle, Alabama: Amer. Ass. Alabama: Geol. Soc. Amer., Bull., Vol. 61, Petrol. Geol., Bull., Vol. 23, p. 1553-1559, p. 1119-1130, 1950. 1939. Clendenin, W. W. A preliminary report upon Stringfield, V. T.; and LaMoreaux, P. E. the Florida parishes of east Louisiana and Age of the Citronelle Formation in Gulf the Bluff Prairie and Hill lands of south- Coastal Plain: Amer. Ass. Petrol. Geol., west Louisiana: in Geology and agriculture Bull., Vol. 41, p. 742-746, 1957. of Louisiana, Part 3, La. State Univ., State Exp. Sta., Geol. Surv., Rep., p. 159- 256, 1896. Doering, John. Citronelle age problem: Amer. Ass. Petrol. Geol., Bull., Vol. 42, p. 1816- MANUSCRIPT RECEIVED BY THE SOCIETY 1862, 1958. OCTOBER 9, 1970

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/82/3/775/3432985/i0016-7606-82-3-775.pdf by guest on 25 September 2021